I will be starting to blog more on data, forecasting, and machine learning related topics here. For starters, I will be doing a short series on the Numerai tournament. For those that have never heard of Numerai, you can check out their website and tournament at https://numer.ai/learn. I also have some code under my machine learning section that can function as a tutorial should you be interested in participating.
The two-line blurb from Numerai describes itself as follows:
Numerai is a data science competition that powers a global equities hedge fund. Top performers are paid from a weekly prize pool.
Using python and the numerapi, you can pull the realized forecast results by the forecasters. I have pulled the data from round 132 to 159, which roughly correspond to November 5, 2018 to May 13, 2019. Numerai has recently changed its forecast performance criteria to AUC (area under curve), and back-calculated the AUC to round 132. Round 159 is the last set of AUC available at the moment.
I have put the data in a 2 panel Tableau Viz. It shows the average live AUC for all submissions on the top, and individual live AUC at the bottom. I will be discussing notable features on the forecasts, the contest, and the behaviors of forecasters over the next set of posts.
